Journal of Experimental Social Psychology xxx (2010) xxx–xxx

ARTICLE IN PRESS

Contents lists available at ScienceDirect

Journal of Experimental Social Psychology

journal homepage: www.elsevier .com/locate / jesp

Thanks, but no thanks: The role of personal responsibility in the experienceof gratitude

Rosalind M. Chow a,*, Brian S. Lowery b

a Carnegie Mellon University, Tepper School of Business, 5000 Forbes Ave., Pittsburgh, PA 15213, USAb Stanford Graduate School of Business, 515 Memorial Way, Stanford, CA 94305 USA

a r t i c l e i n f o

Article history:Received 23 June 2009Revised 10 December 2009Available online xxxx

Keywords:GratitudeResponsibilityResponses to help

0022-1031/$ - see front matter � 2010 Elsevier Inc. Adoi:10.1016/j.jesp.2009.12.018

* Corresponding author.E-mail address: rchow@cmu.edu (R.M. Chow).

Please cite this article in press as: Chow, R. M.,Journal of Experimental Social Psychology (2010)

a b s t r a c t

Current theories of gratitude suggest that individuals feel grateful when they perceive someone else to beresponsible for a desired outcome. However, it is unclear whether individuals must also feel a lack of per-sonal responsibility in order to feel gratitude. This paper provides evidence that in achievement contexts,without the belief that they are responsible for their success, individuals do not experience gratitude,even when they acknowledge the help they have received. In two studies, the more helpful participantsthought an experimenter had been, the more grateful they felt, but only if they also spontaneously feltresponsible for (Study 1) or were induced to feel responsible for (Study 2) their outcomes.

� 2010 Elsevier Inc. All rights reserved.

Introduction

We often depend on others to achieve our goals; students helpeach other to study for tests, coworkers help each other when aproject is not progressing. Although the receipt of such help oftenelicits gratitude (Bartlett & DeSteno, 2006; Lazarus & Lazarus,1994; McCullough, Kilpatrick, Emmons, & Larson, 2001), peopleare not always grateful for help, even if they acknowledge the re-ceipt of help (c.f., Broll, Gross, & Piliavin, 1974; Gross, Wallson, &Paliavin, 1979; Nadler, Fisher, & Streufert, 1974). When and whyis it that people will not feel grateful for help that they recognizehas brought them a desired outcome?

We address these questions by focusing on the foundation ofgratitude: perceptions of responsibility. Although emotion apprai-sal theories assume that the experience of gratitude requires theperception that another person is responsible for a desired out-come (e.g., Weiner, 1979, 1985), it is unclear if individuals mustalso feel a lack of personal responsibility in order to feel gratitude.In this paper, we propose that when outcomes influence how indi-viduals experience the self, such as in achievement contexts, theexperience of gratitude may require both the acknowledgementof help from another and the perceptions of personal responsibil-ity. Conversely, it is possible that individuals may feel less gratefulfor help when they do not feel personally responsible for their out-comes, because the lack of personal responsibility alters theirexperience of the help.

ll rights reserved.

& Lowery, B. S. Thanks, but no, doi:10.1016/j.jesp.2009.12.018

Personal responsibility and gratitude

According to emotion appraisal theories (Roseman, 1984, 1991;Weiner, 1985), gratitude is an attribution-dependent emotion thatarises from a two step process, whereby individuals: (1) recognizethat they have received a positive outcome, and (2) attribute theoutcome to the efforts of another (Weiner, 1985). Thus, gratitudeis fundamentally grounded in perceptions of responsibility; themore individuals perceive others to be responsible for the attain-ment of a desired outcome, the more grateful they should feel(e.g., Solomon, 1977). However, it is unclear if the complementaryassertion – that the more individuals feel responsible for their out-comes, the less grateful they should feel – is true.

Despite this lack of clarity on how gratitude may be influencedby perceptions of personal responsibility, the role of the self hasbeen largely ignored by gratitude researchers. Instead, researchershave focused on how characteristics of the helper or help influ-ences individuals’ propensity to experience gratitude (e.g., Tsang,2006b; Watkins, Scheer, Ovnicek, & Kolts, 2006), and on how theexperience of gratitude can lead to positive psychological and so-cial consequences (e.g., Bartlett & DeSteno, 2006; McCullough, Kil-patrick, Emmons, & Larson, 2001; Tsang, 2006a). In these efforts,researchers have relied on scenario or laboratory inducements ofgratitude where participants can clearly see the effect of otherson their outcomes, but are unlikely to be able to claim personalresponsibility for those outcomes. For example, participants areasked to consider a scenario in which a friend helps them to pur-chase needed textbooks (Tsang, 2006a, 2006b), a friend helps themmove (Watkins et al., 2006), or a confederate helps the participantto complete a study by plugging in a loose monitor cord (Bartlett &

thanks: The role of personal responsibility in the experience of gratitude.

2 R.M. Chow, B.S. Lowery / Journal of Experimental Social Psychology xxx (2010) xxx–xxx

ARTICLE IN PRESS

DeSteno, 2006). In these cases, gratitude is experienced becauseindividuals perceive themselves to have received a benefit fromanother (Solomon, 1977).

However, these situations are also similar in that they do not in-volve outcomes where the individuals’ experience of the self wouldbe influenced by the outcome; whether the outcome is successfulor not holds no import for how the individual feels about the self.In other words, the outcomes studied are not self-relevant. Fromour perspective, in situations where an outcome is not self-rele-vant, individuals are primarily concerned with the outcome itself– whether it is positive or negative – and are unconcerned withwhether they were personally responsible for those outcomes.Thus, in non self-relevant situations, individuals will feel gratefulto the extent that they acknowledge that another person has inten-tionally given, or attempted to give, help that has resulted in a po-sitive outcome (McCullough, Kilpatrick, Emmons, & Larson, 2001;McCullough, Tsang, & Emmons, 2004). This can occur even thoughthey may not feel personally responsible for the outcomes.

In contrast, in many situations, individuals are not only con-cerned with whether they have received a positive outcome or not,they also desire the positive psychological benefits associated witha success, such as positive self-esteem and pride (Weiner, 1979,1985). These types of situations involve self-relevant outcomes, inthat the outcome influences how individuals feel about the self. Insuch situations, personal responsibility becomes a focal concern;an inability to claim responsibility for their outcomes deprives indi-viduals of many of the benefits associated with success (Weiner,1979). To receive a positive outcome without the perception of per-sonal responsibility and its corresponding psychological benefits isnot an outcome for which individuals will feel grateful. Thus, we pro-pose that perceptions of personal responsibility can have a role in theexperience of gratitude, and predict that personal responsibility willplay a positive role in individuals’ experience of gratitude. When out-comes are self-relevant, such as in achievement situations, individ-uals must feel personally responsible for an outcome in order tofeel grateful for help they have received. In contrast, when individu-als do not feel responsible for a success, they will not feel grateful forhelp that led to the positive outcome, even if they acknowledge thatthe help was instrumental to their outcome.

The experience of help

Most attributional theories of emotion assume that feelings ofgratitude are driven by individuals’ experience of the help (e.g.,Lazarus & Lazarus, 1994; Weiner, 1985). According to this perspec-tive, when recipients of help believe that help signifies caring andconcern, they experience the help as supportive (Fisher, DePaulo, &Nadler, 1981; Fisher, Nadler, & Whitcher-Alagna, 1982), whichshould increase their sense of gratitude. Consistent with this theo-retical approach, it is possible that when individuals do not feelresponsible for their success, they underestimate the supportive-ness of a helper. Thus, recipients of help may acknowledge the use-fulness of the help without also experiencing the helper assupportive, which might result in less gratitude.

Overview of studies

We predict that, when assessing an achievement outcome,whether individuals will be grateful for help will depend on: (a)the perceived helpfulness of the help and, (b) the belief they canclaim responsibility for their success. In two studies, we explorethe possibility that when individuals acknowledge that they havebeen helped and feel responsible for their achievements, they willbe more likely to feel grateful to their helper. However, individualswho do not feel responsible for their outcomes will not be gratefulfor help, even if they acknowledge that the help was useful. In

Please cite this article in press as: Chow, R. M., & Lowery, B. S. Thanks, but noJournal of Experimental Social Psychology (2010), doi:10.1016/j.jesp.2009.12.018

Study 2, we also investigate the possibility that this effect is drivenby changes in perceived supportiveness of the help.

Study 1

In Study 1, we examined gratitude towards an experimenterafter the experimenter provided participants with help on anachievement test.

Methods

ParticipantsA total of 52 participants (32 women, 20 men) ranging in age

from 18 to 56 (M = 31.33, SD = 8.82) visited a website containingstudy materials. Participants were recruited from an email listmaintained by a private California university of individuals inter-ested in receiving online survey announcements. As payment, par-ticipants were emailed a $5 gift certificate from an online retailer,and one participant won an additional $5 gift certificate from adrawing. None of the participants reported suspicion about thepurposes or the procedures of the experiment.

ProceduresAfter indicating consent, participants were told that the study

was a two session study about personality and test performance.In the first session, they completed a personality test. After com-pleting the personality test, participants exited the online sessionand were told that they would be emailed information on how toaccess their personality scores. After a day, the experimenteremailed participants a ‘‘personal code” to access their personalityfeedback. In reality, the code gave them access to the experimentwebsite, where they were all given the same personality feedback.

After reading their personality feedback, participants completedan achievement test. To encourage participants to care about theoutcome of the test and to make the test feedback believable, par-ticipants were told that the test was very difficult. Given the diffi-culty of the test, participants who scored above the 90th percentilewould be entered into a drawing for an additional $5 gift card. Be-fore completing the test, participants were told that the experi-menter had the discretion to give some participants a hint. Allparticipants were then shown a hint, which they were led to be-lieve was given due to the experimenter’s familiarity with theirpersonality profile. Thus, participants were led to believe that theexperimenter had chosen to help them in particular. After readingthe hint, participants completed the test.

All participants were then told that they had scored in the 93rdpercentile, a ‘‘high” score, and would therefore be entered into theadditional drawing. They then completed several questionnairesabout their experience, including measures of perceived responsi-bility, helpfulness of the hint, and gratitude. Satisfaction with theirtest performance and positive affect were also measured to controlfor positive affect. After completing the questionnaires, all partici-pants were fully debriefed, and emailed their gift certificates.

Materials and measuresPersonality test. Items on the personality test were taken from therevised Minnesota Multiphasic Personality Inventory (MMPI-2,Butcher, Dahlstrom, Graham, Tellegen, & Kaemmer, 1989). Partici-pants indicated the extent to which they thought 30 statements re-flected who they were as a person. Sample items include, ‘‘I putothers first” and ‘‘I seek out patterns in the universe” (1 = Veryinaccurate, 7 = Very accurate).

Personality feedback. All participants received the same personalityfeedback, which was based on Forer’s personality analysis (also

thanks: The role of personal responsibility in the experience of gratitude.

Table 1Means, standard deviations, and correlations for variables in Study 1.

Variable M SD 1 2 3 4

Responsibility for performance 4.31 .87Perceived helpfulness 4.14 .98 �.05Gratitude 3.64 1.12 .39* .36*

Satisfaction 5.74 1.58 .47** .09 .25Positive affect 3.17 .90 .10 .11 .26 .25

* p < .05.** p < .01.

R.M. Chow, B.S. Lowery / Journal of Experimental Social Psychology xxx (2010) xxx–xxx 3

ARTICLE IN PRESS

called the Barnum effect; Dickson & Kelly, 1985; Forer, 1949). Thefeedback is supposedly tailored to individuals, but is in fact vagueand general enough to apply to a large number of people. Somesample statements from the feedback include: ‘‘You have a needfor other people to like and admire you, and yet you tend to be crit-ical of yourself. While you have some personality weaknesses youare generally able to compensate for them” and ‘‘Disciplined andself-controlled on the outside, you tend to be worrisome and inse-cure on the inside.”

After reading their feedback, participants were asked to rate theaccuracy of the feedback (1 = extremely inaccurate, 4 = neitheraccurate nor inaccurate, 7 = extremely accurate). Participants per-ceived the feedback to be fairly accurate, M = 5.21, SD = 1.27.

Achievement test. After receiving their personality feedback, partic-ipants were asked to complete a 12 item test of ‘‘integrative orien-tation.” Integrative orientation was described as being ‘‘related toone’s ability to see solutions to problems and to solve problemscreatively.” In reality, the test was the Remotes Associates Task(RAT, Mednick, 1962). Each RAT item consists of three stimuluswords that are related to a fourth unreported word, which is gen-erated by the participant. For example, an item might consist of thestimulus words: ‘‘falling”, ‘‘actor”, and ‘‘dust.” A correct responsewould be the fourth word ‘‘star.”1

Help. All participants were told that the experimenter wanted togive them a hint for the test, based on their personality test results.Specifically, participants read a note from the experimenter, whichstated, ‘‘Based on your personality profile, I want to give you a hintfor the test – all of the solutions to the following problems areeither colors or involve nature (plants, animals, etc.).” Items onthe test were chosen to have answers that corresponded only tocolors or plants and animals. Thus, the hint narrowed the numbercategories that participants had to consider, but still allowed par-ticipants to feel that they had generated the answer on their own.

Responsibility. To measure how responsible participants felt fortheir performance on the test, participants were asked ‘‘Howresponsible do you feel for your performance on this test?”(1 = Not at all responsible, 5 = extremely responsible).

Perceived helpfulness. To measure how helpful participants thoughtthe experimenter had been, they were asked ‘‘How helpful was theexperimenter?” (1 = Not at all, 5 = extremely).

Gratitude. Participants were asked how grateful they felt towardsthe experimenter (1 = Not at all, 5 = extremely).

Satisfaction with outcome. Participants were asked to rate how sat-isfied they were with their test performance (1 = extremely unsat-isfied, 4 = neutral, 7 = extremely satisfied).

Positive affect. Participants were asked to rate the extent to whichthey felt ten positive emotions (e.g., proud, strong, and excited;1 = not at all, 5 = extremely, a = .94). Emotion terms were thoseused in the Positive and Negative Affect Scale (PANAS, Watson,Clark, & Tellegen, 1988).

Results

Means, standard deviations, and correlations are shown inTable 1.

1 Participants averaged 7.90 correct answers (SD = 2.77) out of 12, suggesting theyshould have found the high test performance feedback believable.

Please cite this article in press as: Chow, R. M., & Lowery, B. S. Thanks, but noJournal of Experimental Social Psychology (2010), doi:10.1016/j.jesp.2009.12.018

We predicted that the more helpful participants thought theexperimenter had been, the more grateful they would feel, but onlyif they also felt responsible for their test performance. To test thishypothesis, we mean centered participants’ ratings of perceivedresponsibility and helpfulness of the experimenter, and then mul-tiplied them to create an interaction term (cf. Aiken & West, 1991).We then regressed participants’ feelings of gratitude on howresponsible they felt for their test performance, how helpful theythought the experimenter had been, and the interaction term. Sat-isfaction and positive affect were entered as control variables.2

The more helpful participants perceived the hint to be, the moregratitude they reported, B = .39, SE B = .15, b = .36, t(46) = 2.55,p < .05. Perceived responsibility for test performance did not affectgratitude, B = .01, SE B = .12, b = .01, t(46) = .06, p = .95. Impor-tantly, we observed the predicted Perceived responsibility � Help-fulness interaction on gratitude, B = .34, SE B = .14, b = .31,t(46) = 2.36, p < .05, see Fig. 1. This effect was not moderated byparticipant gender.

Simple slopes analyses (Aiken & West, 1991) revealed that, aspredicted, among individuals who felt responsible for their out-comes (+1 SD), the more they thought the experimenter hadhelped them, the more grateful they reported feeling, B = .75, SEB = .22, b = .69, t(46) = 3.43, p = .001. In contrast, among individualswho felt less responsible (�1 SD), increases in the acknowledgedhelpfulness of the hint was not associated with greater gratitude,B = .03, SE B = .21, b = .02, t(46) = 2.36, p = .90.

Discussion

Study 1 provided evidence for the hypothesis that how gratefulindividuals feel for help depends not only on how much help theythink they have received, but also on whether they feel responsiblefor their outcome. Increases in perceived helpfulness were onlyassociated with more gratitude when individuals felt responsiblefor their outcomes. Individuals who did not feel responsible fortheir outcomes were not more grateful to the experimenter, evenif they acknowledged the helpfulness of the hint.

Although the results of Study 1 support our hypotheses, thestudy relied on measured variables, and thus cannot establish cau-sality. Study 2 was designed to address this issue and to explorethe possibility that the effect is mediated by the perceived suppor-tiveness of the help.

Study 2

Method

ParticipantsA total of 60 participants (37 women, 23 men) ranging in age

from 18 to 55 (M = 31.93, SD = 9.99) visited a website containing

2 Removal of satisfaction and positive affect as covariates did not reduce thesignificance of the Perceived responsibility � Perceived helpfulness interaction ongratitude, B = .33, SE B = .14, b = .31, p < .05.

thanks: The role of personal responsibility in the experience of gratitude.

1

2

3

4

5

HighLow

Gra

titud

e

Perceived Helpfulness of Experimenter

Low perceived responsibility

High perceived responsibility

Fig. 1. Gratitude as a function of perceived responsibility for test performance andperceived helpfulness of experimenter hint (Study 1).

4 R.M. Chow, B.S. Lowery / Journal of Experimental Social Psychology xxx (2010) xxx–xxx

ARTICLE IN PRESS

study materials. Participants were recruited from an email listmaintained by a private California university of individuals inter-ested in receiving online survey announcements. As payment, par-ticipants were emailed a $7 gift certificate from an online retailer.None of the participants reported suspicion about the purposes orthe procedures of the experiment.

ProcedureThe procedures were similar to those used in Study 1. Partici-

pants were given a personality test, and were given bogus feedbackon their personality.3 Participants were then told that the next partof the study was designed to simulate the experience of a manager,and that the test was designed to measure individuals’ managerialcapability. To ensure that high performance on the test was per-ceived to be a positive outcome, participants were told that theywould have the opportunity to earn an extra $2 (in addition to anadvertised $5 payment) based on their performance on the test. Be-fore participants completed the test, they were provided with thesame hint used in Study 1, ostensibly due to the experimenter’sfamiliarity with their personality profile.

Participants then took the integrative orientation test used inStudy 1, and all were told that they had scored well enough to re-ceive the additional payment. Participants also received feedbackon their performance that indicated that they either were or werenot responsible for this outcome. After receiving their scores andfeedback, participants completed a series of questionnaires ontheir experience, and then were debriefed and emailed their giftcertificates.

4 Participants averaged 7.60 correct answers (SD = 3.18).5 It is possible that the effect of our responsibility manipulation on gratitude was

driven by either participants’ perceptions of their performance or the importance theyplaced on receiving the bonus. To test these possibilities, we replicated Study 2 with41 participants (24 women, 17 men) and added measures of perceived performanceand the importance of the bonus. Perceived performance was measured by the item,‘‘I performed well on the Integrative Orientation Task,” (1 = strongly disagree,7 = strongly agree) and importance of the bonus was measured by the item, ‘‘Howimportant was receiving a bonus to you?” (1 = Not at all important, 5 = extremelyimportant). Perceived performance did not differ across the Not responsible, M = 4.94,SD = 1.35, and Responsible conditions, M = 5.38, SD = 1.53, t(39) = .94, p = .35, which isinconsistent with the possibility that participants’ decreased levels of gratitude weredriven by their perceptions of their performance. Moreover, participants in the Not

Materials and manipulationsTask description. All participants were told that in addition to themanagerial capability test, the study would also include three fac-tors that would simulate different managerial environments: a bo-nus, having a mentor, and the element of chance. To simulate theexperience of ‘‘pay for performance,” all participants could ‘‘earna bonus” of $2 if they scored above 90% in the simulation. In addi-tion, to simulate advice giving from mentors, participants weretold that the experimenter could choose to give some participantsa hint (in reality, all participants were given a hint). Finally, thetask also included a simulation of chance, in that participants’ finalscores (on which their bonus depended) would be affected by ran-

3 On average, the feedback was perceived to be accurate (1 = extremely inaccurate,4 = neither accurate nor inaccurate, 7 = extremely accurate), M = 5.41, SD = 1.28.

Please cite this article in press as: Chow, R. M., & Lowery, B. S. Thanks, but noJournal of Experimental Social Psychology (2010), doi:10.1016/j.jesp.2009.12.018

dom chance. Participants were told that after taking the test, thecomputer would generate a percentage from �10% to 10% thatwould be added to the participants’ test score in order to deter-mine their final score. For example, if a participant scored 88% cor-rect on the test, and the computer chose +5%, the participants’ finalscore would be 93%, resulting in the bonus.

Help. All participants received the same hint as in Study 1.

Responsibility manipulation. After completing their test, partici-pants were given their test scores.4 Participants in the Responsiblecondition were shown:

Your managerial simulation results:

– Percentage answers correct on the Integrative Orientation Task:97%.

– Percentage chosen by computer: �5%.– Final managerial performance score: 92%.

You will get a bonus. Based on your performance, if you were amanager, you would be directly responsible for the success of yourproject or team.

Participants in the Not responsible condition were shown:Your managerial simulation results:

– Percentage answers correct on the Integrative Orientation Task:87%.

– Percentage chosen by computer: +5%.– Final managerial performance score: 92%.

You will get a bonus. Based on your performance, if you were amanager, you would not be directly responsible for the success ofyour project or team.5

Manipulation check. To ensure our manipulation of responsibilityworked as intended, we asked participants, ‘‘How responsible doyou feel you were for your bonus?” (1 = Not at all responsible,5 = extremely responsible).

Perceived helpfulness. This measure was identical to the one used inStudy 1.

Gratitude. Because our manipulation of responsibility involved anelement of chance, we wanted to ensure that our measure of grat-itude referred specifically to the experimenter and the hint. There-fore, we asked: ‘‘To what extent did getting the hint make you feelgrateful?” (1 = Not at all, 5 = extremely) and ‘‘To what extent doyou feel grateful to the experimenter for helping you?” (1 = Notat all grateful, 5 = extremely grateful, a = .89).

responsible condition did not rate receiving the bonus as less important, M = 3.29,SD = 1.16, than participants in the Responsible condition, M = 3.75, SD = .90,t(39) = 1.42, p = .16, suggesting that participants did not defensively downplay theimportance of the bonus in response to not being responsible for the bonus.

thanks: The role of personal responsibility in the experience of gratitude.

Table 2Means, standard deviations, and correlations for variables in Study 2.

Variable M SD 1 2 3 4

Perceived helpfulness 3.97 .95Experienced support 4.19 .91 .36**

Gratitude 4.09 .90 .48** .73**

Satisfaction 5.60 1.51 .31* .34** .28*

Positive affect 3.27 .92 .25* .28* .29* .55**

* p < .05.** p < .01.

Table 3Independent sample t-tests across responsibility conditions for variables in Study 2.

Variable Condition

Not responsible ResponsibleMean (SD) Mean (SD) t(df)

Responsibility for performance 2.56 (1.22) 3.64 (1.27) 3.34 (61)**

Perceived helpfulness 4.07 (.90) 3.89 (.99) .77 (61)Gratitude 4.09 (.81) 4.09 (.98) .02 (61)Experienced support 3.32 (.88) 3.61 (.76) 1.42 (61)Satisfaction 5.11 (1.70) 6.00 (1.23) 2.36 (61)*

Positive affect 3.27 (1.01) 3.30 (.86) .13 (61)

* p < .05.** p < .01.

1

2

3

4

5

Gra

titud

e

Not responsible

Responsible

R.M. Chow, B.S. Lowery / Journal of Experimental Social Psychology xxx (2010) xxx–xxx 5

ARTICLE IN PRESS

Experienced support. To measure participants’ perceptions of sup-port, they rated the extent to which they thought the experimenterhad been encouraging, supportive, condescending, dismissive, andpatronizing (the last three items were reverse scored, 1 = Not at all,5 = extremely, a = .77).

Satisfaction with outcome and positive affect. Measures of satisfac-tion and positive affect were identical to those used in Study 1.

Results and discussion

Preliminary analysesMeans, standard deviations, and correlations are shown in

Table 2.To ensure our manipulation of responsibility worked as in-

tended, we ran an independent samples t-test on participants’ feel-ings of responsibility across the Responsible and Not responsibleconditions. Participants in the Not responsible condition felt lessresponsible for their bonus than participants in the Responsiblecondition. Moreover, a one-sample t-test indicated that perceivedresponsibility was significantly higher than the mid-point of threeamong participants in the Responsible condition, t(32) = 2.88,p < .01.

It is possible that participants would perceive the experimenterto be less helpful, and therefore be less grateful, in the Not Respon-sible condition than in the Responsible condition, since their ‘‘suc-cess” was due to chance, and not due to the experimenter’s help. Totest this possibility, we also conducted an independent samples t-test on perceived helpfulness. Participants did not differ in theirratings of the helpfulness of the hint and experimenter acrossresponsibility conditions, suggesting that differences in gratitudewere not due to participants’ perceptions of the help. Independentsample t-test results for all study variables across Responsibilityconditions are shown in Table 3.

Main analysesGratitude. We hypothesized that the more helpful participantsthought the experimenter had been, the more grateful they wouldfeel, but only if they were led to believe they were responsible fortheir success. To test this interactive hypothesis, we dummy codedthe Responsibility manipulation (�1 = Not responsible, 1 = Respon-sible), mean-centered perceived helpfulness, and computed theirinteraction term. We then regressed participants’ levels of grati-tude on Responsibility condition, perceived helpfulness, and theirinteraction term. Satisfaction and positive affect were entered ascontrol variables.6

The more helpful participants thought the experimenter hadbeen, the more grateful they reported feeling, B = .35, SE B = .11,b = .37, t(54) = 3.09, p < .005. Levels of gratitude were not affected

6 Removal of satisfaction and positive affect as covariates did not reduce thesignificance of the Perceived responsibility � Perceived helpfulness interaction ongratitude, B = .25, SE B = .11, b = .26, p < .05, or experienced support, B = .29, SE B = .11,b = .34, p < .01.

Please cite this article in press as: Chow, R. M., & Lowery, B. S. Thanks, but noJournal of Experimental Social Psychology (2010), doi:10.1016/j.jesp.2009.12.018

by the Responsibility manipulation, B = �.03, SE B = .11, b = �.03,t(54) = �.24, p = .80. Importantly, replicating the results of Study1, there was a significant interaction between responsibility andperceived helpfulness on gratitude, B = .27, SE B = .11, b = .28,t(54) = 2.52, p < .05, see Fig. 2. This effect was not moderated byparticipant gender.

Closer examination of the interaction revealed that among indi-viduals in the Responsible condition, the more participantsthought the experimenter had been helpful, the more grateful theyfelt, B = .62, SE B = .14, b = .66, t(54) = 4.53, p < .001. In contrast,individuals in the Not responsible condition were not more grate-ful for help, even as the perceived helpfulness of the experimenterincreased, B = .08, SE B = .17, b = .09, t(54) = .48, p = .63.

Experienced support. To test the possibility that help is experiencedas more supportive when individuals feel responsible for their out-comes than when they do not, we reran the above analyses, replac-ing gratitude with participants’ experience of support. Participantratings of experimenter supportiveness were not affected by howhelpful the hint had been, B = .11, SE B = .10, b = .15, t(54) = 1.11,p = .27. However, participants in the Responsible condition tendedto rate the experimenter as more supportive than participants inthe Not Responsible condition, B = .16, SE B = .09, b = .23,t(54) = 1.79, p = .08. Importantly, there was a significant Responsi-bility manipulation � Perceived helpfulness interaction on theexperience of experimenter support, B = .21, SE B = .09, b = .28,t(54) = 2.31, p < .05, see Fig. 3.

Inspection of the interaction revealed that among participantsin the Responsible condition, the more helpful the experimenter

hgiH owLPerceived Helpfulness of Experimenter

Fig. 2. Gratitude as a function of Responsibility condition and perceived helpfulnessof experimenter (Study 2).

thanks: The role of personal responsibility in the experience of gratitude.

1

2

3

4

5

High Low

Expe

rienc

ed S

uppo

rt

Perceived Helpfulness of Experimenter

Not responsible

Responsible

Fig. 3. Experienced support as a function of Responsibility condition and perceivedhelpfulness of experimenter (Study 2).

6 R.M. Chow, B.S. Lowery / Journal of Experimental Social Psychology xxx (2010) xxx–xxx

ARTICLE IN PRESS

was perceived to have been, the more participants thought theexperimenter had been supportive, B = .31, SE B = .12, b = .43,t(54) = 2.72, p < .01. In contrast, participants in the Not responsiblecondition did not differ in how they rated the experience of thehelp, even as they acknowledged the helpfulness of the experi-menter, B = �.10, SE B = .15, b = �.14, t(54) = �.69, p = .49.

Mediational analyses. To explore the possibility that supportive-ness mediates the interactive effect of help and responsibility, weconducted a moderated mediation model (cf. Preacher, Rucker, &Hayes, 2007). This model tests the hypothesis that perceived sup-portiveness mediates the effect of help in the responsible condi-tion, but not in the Not responsible condition, and consists ofthree regression analyses. The first two regressions were identicalto those described above; the finding that the Responsibility condi-tion � Perceived helpfulness interaction predicts both gratitudeand the experience of support is consistent with our mediationalhypothesis.

To test the final step of the mediation, we regressed gratitudeon Responsibility condition, perceived helpfulness, the Responsi-bility manipulation � perceived helpfulness interaction, control-ling for the effects of experienced supportiveness. The analysisrevealed that controlling for the experience of support reducedthe effect of the Responsibility condition � perceived helpfulnessinteraction on gratitude to nonsignificance, B = .07, SE B = .09,b = .08, p = .44. We also examined the conditional indirect effectsamong participants in the Responsible and Not responsible condi-tions. These effects indicated that experienced support mediatedthe relationship between perceived helpfulness and gratitudeamong participants in the Responsible condition, z = 2.83,p < .005, but not among participants in the Not responsible condi-tion, z = .86, p = .39.

Although existing theory suggests that perceived supportive-ness drives gratitude, it is possible that the experience of suppor-tiveness is driven by individuals’ gratitude. To test thispossibility, we reran the meditational analyses, reversing the posi-tions of gratitude and experienced support. Examination of theconditional indirect effects revealed that, among participants inthe Responsible condition, levels of gratitude mediated the rela-tionship between perceived help and experienced support,z = 3.79, p < .001, but this was not true among participants in theNot responsible condition, z = 1.02, p = .31. Thus, our data do notprovide unambiguous evidence that supportiveness mediates theinteractive effect of help and responsibility on gratitude. We

Please cite this article in press as: Chow, R. M., & Lowery, B. S. Thanks, but noJournal of Experimental Social Psychology (2010), doi:10.1016/j.jesp.2009.12.018

encourage future research to determine whether emotional re-sponses to help are driven by the experience of help, or if the expe-rience of help is influenced by individuals’ emotional responses tohelp.

General discussion

Current theories of emotion assume that individuals must be-lieve others to be responsible for a desired outcome in order to feelgrateful (Weiner, 1985; Weiner, Russell, & Lerman, 1978, 1979).However, this leaves open the question of whether individualsmust also believe that they are not personally responsible for theiroutcomes in order to feel gratitude. These studies suggest thatwhen outcomes are self-relevant, as in achievement situations,feelings of personal responsibility for outcomes can enhance indi-viduals’ gratitude towards others when they also acknowledge thereceipt of help.

Two studies demonstrated that acknowledgement of help, com-bined with feelings of personal responsibility for outcomes, in-creases individuals’ propensity to feel gratitude in achievementsituations. The more helpful participants thought an experimenterhad been, the more grateful they felt towards the experimenter,but only if they felt responsible for their test performance (Study1) or were led to believe they were responsible for their test perfor-mance (Study 2). In Study 2, participants’ levels of gratitude werelinked to experienced support, suggesting that when individualsdo not feel responsible for their outcomes, they are less likely toexperience help as supportive – even though they acknowledgethe value of the help received. However, we were unable to deter-mine whether the belief that a helper is supportive increases grat-itude or if increased gratitude leads to perceptions of greatersupportiveness. Further research is necessary to more clearlyestablish the direction of this relationship.

The experience of responsibility

Our findings suggest that the acknowledgement of help neednot preclude the experience of personal responsibility (and viceversa). It is possible for individuals to continue to claim responsi-bility for an outcome while acknowledging that others have helpedthem. Consistent with this possibility, measured perceptions ofpersonal responsibility were not correlated with perceptions ofhelpfulness (Study 1) and manipulated perceptions of personalresponsibility did not affect individuals’ assessment of helpfulness(Study 2). These findings dovetail with theory that suggests thatinternal and external attributions can be independent (McClure,1991, 1998). The present results are also consistent with thehypothesis that grateful individuals can have an expanded circleof attributions, such that they attribute their success to others’ ac-tions, but also take into account how they themselves have con-tributed to their own success (McCullough, Emmons, & Tsang,2002).

One question that arises from this research is why our resultsdiffered from those typically found by emotion researchers (e.g.,Weiner, 1979; Weiner et al., 1978). We believe that previous re-search methods, which directly manipulated participants’ apprais-als for events or used bipolar attribution scales, constrainedresearchers’ abilities to detect the independence of personal andothers’ responsibility for an event. These approaches force internaland external attributions to be mutually exclusive. In our studies,however, we measured internal and external attributionsseparately.

Interestingly, individuals’ experience of help seems to be lessinfluenced by why they are or are not responsible for an outcomethan by whether or not they are responsible. For example, in Study

thanks: The role of personal responsibility in the experience of gratitude.

R.M. Chow, B.S. Lowery / Journal of Experimental Social Psychology xxx (2010) xxx–xxx 7

ARTICLE IN PRESS

2, participants were told that although they would receive the po-sitive outcome, they were not responsible for their outcomes be-cause it was attained through chance. Thus, it was not the helpthat deprived them of responsibility for their outcome; indeed,all participants received the same amount of help. Reflecting this,participants did not rate the experimenter’s helpfulness differentlyacross conditions. Yet the lack of responsibility led them to experi-ence the experimenter as less supportive, even though the experi-menter was perceived to be equally helpful. Thus, our resultssuggest that simply feeling responsible for an outcome is sufficientto change the experience of help.

The experience of help

Research on when individuals will feel gratitude has often fo-cused on what characteristics of the help or of the helper elicitgratitude. For example, the experience of gratitude is influencedby whether the help is perceived to be valuable, costly to the ben-efactor, given with good intentions, and given gratuitously (Bar-Tal, Bar-Zohar, Greenberg, & Hermon, 1977; Graham, 1988; Lane& Anderson, 1976; Tesser, Gatewood, & Driver, 1968). In contrast,our findings suggest that the experience of help is not simply afunction of the help or the helper, but instead also depends onhow individuals experience the self. For example, participants’experience of the helper’s supportiveness was jointly determinedby the actual amount of help given by the experimenter and theirperception of their responsibility for an outcome.

We believe that this difference may stem from previous grati-tude researchers’ tendency to study gratitude within non-achieve-ment contexts, where the outcomes are less self-relevant. Incontrast, our findings align with literature that emphasizes thatfactors other than help itself can play a role in the way individualsexperience help (see Fisher et al., 1982 for a review). This line ofresearch was largely conducted within the achievement domain.Thus, the present research suggests two possibilities: (1) that togain a fuller understanding of individuals’ experience of gratitude,we must not only consider helper and help characteristics, but alsohow the experience of the self influences individuals’ propensity tofeel grateful, and (2) that one fruitful way to study the effects ofothers and the self is to study gratitude in both achievement andnon-achievement contexts.

Acknowledgments

The authors would like to thank Caitlin Hogan, Lara Kammrath,Dale Miller, Elizabeth Mullen, and Larissa Tiedens for constructivecomments on drafts of this manuscript.

References

Aiken, L. S., & West, S. G. (1991). Multiple regression: Testing and interpretinginteractions. Newbury Park, CA: Sage.

Bar-Tal, D., Bar-Zohar, Y., Greenberg, M. S., & Hermon, M. (1977). Reciprocitybehavior in the relationship between donor and recipient and between harm-doer and victim. Sociometry, 40, 293–298.

Bartlett, M. Y., & DeSteno, D. (2006). Gratitude and prosocial behavior: Helpingwhen it costs you. Psychological Science, 17, 319–325.

Please cite this article in press as: Chow, R. M., & Lowery, B. S. Thanks, but noJournal of Experimental Social Psychology (2010), doi:10.1016/j.jesp.2009.12.018

Broll, L., Gross, A. E., & Piliavin, I. (1974). Effects of offered and requested help onhelp-seeking and reactions to being helped. Journal of Applied Social Psychology,4, 244–358.

Butcher, J. N., Dahlstrom, W. G., Graham, J. R., Tellegen, A., & Kaemmer, B. (1989).Manual for the restandardized Minnesota Multiphasic Personality Inventory:MMPI-2: An administrative and interpretive guide. Minneapolis: University ofMinnesota Press.

Dickson, D. H., & Kelly, I. W. (1985). The Barnum effect in personality assessment: Areview of the literature. Psychological Reports, 57, 367–382.

Fisher, J. D., DePaulo, B. M., & Nadler, A. (1981). Extending altruism beyond thealtruistic act: The mixed effects of aid on the help recipient. In A. Nadler, J. D.Fisher, & B. M. DePaulo (Eds.). New directions in helping (Vol. 3, pp. 163–185).New York: Academic Press.

Fisher, J. D., Nadler, A., & Whitcher-Alagna, S. (1982). Recipient reactions to aid.Psychological Bulletin, 91, 27–54.

Forer, B. R. (1949). The fallacy of personal validation: A classroom demonstration ofgullibility. Journal of Abnormal Psychology, 44, 118–121.

Graham, S. (1988). Children’s developing understanding of the motivational role ofaffect: An attributional analysis. Cognitive Development, 3, 71–88.

Gross, A. E., Wallson, B. S., & Paliavin, I. (1979). Reactance, attributions, equity, andthe help recipient. Journal of Applied Social Psychology, 9, 297–313.

Lane, J., & Anderson, N. H. (1976). Integration of intention and outcome in moraljudgment. Memory and Cognition, 4, 1–5.

Lazarus, R. S., & Lazarus, B. N. (1994). Passion and reason: Making sense of ouremotions. New York: Oxford University Press.

McClure, J. L. (1991). Explanations, accounts, and illusions: A critical analysis.Cambridge: Cambridge University Press.

McClure, J. L. (1998). Discounting causes of behavior: Are two reasons better thanone? Journal of Personality and Social Psychology, 74, 7–20.

McCullough, M. E., Emmons, R. A., & Tsang, J. (2002). The grateful disposition: Aconceptual and empirical topography. Journal of Personality and SocialPsychology, 82, 112–127.

McCullough, M. E., Kilpatrick, S. D., Emmons, R. A., & Larson, D. B. (2001). Is gratitudea moral affect? Psychological Bulletin, 127, 249–266.

McCullough, M. E., Tsang, J., & Emmons, R. A. (2004). Gratitude in ‘‘intermediateaffective terrain”: Grateful moods and their links to personality and daily lifeevents. Journal of Personality and Social Psychology, 86, 295–309.

Mednick, S. A. (1962). The associative basis of the creative process. PsychologicalReview, 69, 220–232.

Nadler, A., Fisher, J. D., & Streufert, S. (1974). The donor’s dilemma: Recipient’sreactions to aid from friend or foe. Journal of Applied Social Psychology, 4,275–285.

Preacher, K., Rucker, D., & Hayes, A. (2007). Assessing moderated mediationhypotheses: Theory, methods, and prescriptions. Multivariate BehavioralResearch, 42, 185–227.

Roseman, I. J. (1984). Cognitive determinants of emotion: A structural theory. In P.Shaver (Ed.). Review of personality and social psychology: Emotions, relationships,and health (Vol. 5, pp. 11–36). Beverly Hills, CA: Sage.

Roseman, I. J. (1991). Appraisal determinants of discrete emotions. Cognition andEmotion, 25, 745–751.

Solomon, R. C. (1977). The passions. Garden City, NY: Anchor Books.Tesser, A., Gatewood, R., & Driver, M. (1968). Some determinants of gratitude.

Journal of Personality and Social Psychology, 9, 233–236.Tsang, J. (2006a). Gratitude and prosocial behavior: An experimental test of

gratitude. Cognition and Emotion, 20, 138–148.Tsang, J. (2006b). The effects of helper intention on gratitude and indebtedness.

Motivation and Emotion, 30, 199–205.Watkins, P. C., Scheer, J., Ovnicek, M., & Kolts, R. (2006). The debt of gratitude:

Dissociating gratitude and indebtedness. Cognition and Emotion, 20, 217–241.Watson, D., Clark, L. A., & Tellegen, A. (1988). Development and validation of brief

measures of positive and negative affect: The PANAS scales. Journal ofPersonality and Social Psychology, 54, 1063–1070.

Weiner, B. (1979). A theory of motivation for some classroom experiences. Journal ofEducational Psychology, 71, 3–25.

Weiner, B. (1985). An attributional theory of achievement motivation and emotion.Psychological Review, 92, 548–573.

Weiner, B., Russell, D., & Lerman, D. (1978). Affective consequences of causalascriptions. In J. H. Harvey, W. J. Ickes, & R. F. Kidd (Eds.). New directions inattribution research (Vol. 2). Hillsdale, NJ: Erlbaum.

Weiner, B., Russell, D., & Lerman, D. (1979). The cognition–emotion process inachievement related contexts. Journal of Personality and Social Psychology, 37,1211–1220.

thanks: The role of personal responsibility in the experience of gratitude.